2  Data processing in TIMSImaging

2.1 Introcution

In this part, we start from raw .d data, explore the dataset, perform peak processing, show separation by ion mobility and export the results in the open imzML format.

import timsimaging

# enable visualization in the Jupyter notebook
from bokeh.io import show, output_notebook
output_notebook()
# disable FutureWarning
import warnings
warnings.filterwarnings('ignore')
Loading BokehJS ...

2.1.1 Load MALDI-TIMS-TOF raw data

Suppose we know nothing about this dataset. First we load the dataset and print basic information like pixel number and data range.

bruker_d_folder_name = r"D:\dataset\Kidney_MS1_ITO6.d"
dataset = timsimaging.spectrum.MSIDataset(bruker_d_folder_name)
dataset
MSIDataset with 38267 pixels
        mz range: 799.998-2500.007
        mobility range: 1.200-2.600

2.1.2 View summary image and tables

Then we can view the TIC image without any processing, as the TIC intensities are stored in the Frames table in the raw tdf file. Hover on individual pixels to see the details like coordinates and intensity:

dataset.image()

We can view the underlying data as tables, too

# pixel positions
dataset.pos
XIndexPos YIndexPos
Frame
1 336 175
2 337 175
3 338 175
4 339 175
5 340 175
... ... ...
38263 737 116
38264 738 116
38265 739 116
38266 740 116
38267 741 116

38267 rows × 2 columns

# metadata of pixels
dataset.data.frames
Id Time ScanMode MsMsType TimsId MaxIntensity SummedIntensities NumScans NumPeaks MzCalibration T1 T2 TimsCalibration PropertyGroup AccumulationTime RampTime Pressure Denoised
0 0 0.000000 20 0 64 0 0 1212 0 1 25.630815 23.592239 1 1 59.47 230.09 2.055982 0
1 1 2.671587 20 0 64 688 5346970 1212 76165 1 25.630815 23.592239 1 1 59.47 230.09 2.055982 0
2 2 2.941060 20 0 210641 637 4998845 1212 71095 1 25.630815 23.592239 1 1 59.47 230.09 2.055982 0
3 3 3.210677 20 0 407604 605 4793642 1212 67948 1 25.630815 23.592239 1 1 59.47 230.09 2.055982 0
4 4 3.480285 20 0 595229 626 4705085 1212 66534 1 25.630938 23.591317 1 1 59.47 230.09 2.055983 0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
38263 38263 10602.755832 20 0 11938087511 925 4587513 1212 64378 1 25.666533 23.497069 1 1 59.47 230.09 2.055988 0
38264 38264 10603.025466 20 0 11938264725 720 4511870 1212 63215 1 25.666508 23.496083 1 1 59.47 230.09 2.055988 0
38265 38265 10603.295060 20 0 11938440038 775 4700548 1212 66076 1 25.666508 23.496083 1 1 59.47 230.09 2.055988 0
38266 38266 10603.564667 20 0 11938622040 626 4648396 1212 65487 1 25.666508 23.496083 1 1 59.47 230.09 2.055988 0
38267 38267 10603.813145 20 0 11938802286 755 4574022 1212 63950 1 25.666508 23.496083 1 1 59.47 230.09 2.055988 0

38268 rows × 18 columns

2.1.3 Set region of interest(ROI)

The left region is the kidney tissue, we can use the set_ROI method to set the ROI by xmin, xmax, ymin, ymax parameters.

dataset.set_ROI(name='kidney', xmax=500)

2.1.4 Process selected ROI

Then we run the processing pipline introduced in the basic use notebook.

results = dataset.process(sampling_ratio=0.1, frequency_threshold=0.02, tolerance=3, adaptive_window=True, roi='kidney', visualize=True)
Computing mean spectrum...
Traversing graph...
Finding local maxima...
Summarizing...

2.1.5 View peak list

We can view the interactive peak list:

peaklist = results["peak_list"]
table, _ = timsimaging.plotting.feature_list(results["peak_list"])
show(table)

2.1.6 View intensity distribution

We can also visualize the intensity distribution:

results["peak_list"]["total_intensity"].hist()

2.1.7 View Interactive Results

User can explore the results in the GUI. By clicking pixels in the image(top left), users can view its mass spectrum(bottom middle), mobilogram(top right) and 2D spectogram combining these two(top middle). Also, users can click entries in the peak list(bottom left) to investigate ion images and spectral details of that peak. The menu(bottom right) provide summarizing functions like TIC image and mean spectrum, as well as file export options. This requires a living Python session to render, here shows a screenshot:

show(results['viz'])

2.1.8 Separate ions by ion mobility

Isobaric ions, or ions with non-distinguishable m/z(for example, isomers) could be separated by ion mobility. As examples shown below, TIMSImaging detected both peaks.

from bokeh.layouts import row
isomer1, _  = timsimaging.plotting.image(dataset, i=1079, results=results)
isomer2, _  = timsimaging.plotting.image(dataset, i=1080, results=results)
show(row(isomer1, isomer2))

isomer1, _  = timsimaging.plotting.image(dataset, i=1163, results=results)
isomer2, _  = timsimaging.plotting.image(dataset, i=1164, results=results)
show(row(isomer1, isomer2))

2.1.9 Export the processed data in imzML format

Finally, we export the processed data as continuous, centroid imzML file for downstream analysis in Cardinal.

timsimaging.spectrum.export_imzML(dataset, path=r"D:/dataset/Kidney_MS1_ITO6", peaks=results)